dmed (David Medina)
dmed (David Medina)
Implement an initial `array.sort()` method based on the backend: - CPU: `std::sort` - GPU: bitonic sort Eventually would be nice to have more performant kernels based on the type (e.g....
There is currently a ton of logic to decide whether an operator is left unary or not. The check should be simpler than it is at the moment.
There are currently about 150 `OCCA_ERROR`s and 150 `printError`s, it would be nice to standardize an error code format and make it easy to document and find reasons/causes For example,...
Right now `fileOrigin` has a push that takes a `fromInclude` bool, we should split that into 2 methods to make it cleaner
The exclusive index variable increments in the inner-most `@inner` loop and is reset to 0 in the inner-most `@outer`-loop In reality, we should be resetting the index in the `@inner`...
`@exclusive` array sizes for CPU modes are hard-coded to `256` We should use allocate the size depending on the full `@inner` loop size
We're storing file contents everytime there is an `#include`, as opposed to caching the file source
Use appropriate backend locks to support: - [ ] Update expressions (`+=, -=, *=, /=` updates) - [ ] Block statements (critical regions) ### Atomic expressions #### Inlined atomic ```cpp...
Test out `Serial` and `OpenMP` modes with ARM hardware to make sure compiler and compiler flags work. Find ARM-specific macros for vectorization options (`float2`, `float4`, ...)
It would be nice to add proper indentation and highlighting to OKL attributes for - [x] Emacs (added [0a776961f420e0eb7c46d8cefdc35644a93b43e0]) - [ ] Vi/Vim - [ ] Kate/Qt Creator - [...